Advances in the diagnosis of fungal pneumonias

2.1.1. Culture and Direct Visualization

The gold standard for diagnosing IPA is direct visualization of angioinvasion by acute angle, branching septate hyphae on histopathology with a positive Aspergillus spp. culture (Table 1) [6]. Unfortunately histopathological diagnosis is often not feasible secondary to illness severity and bleeding risk. Therefore, other means have been used to help establish diagnosis.

Sputum culture with Aspergillus spp. growth has different implications depending on host risk factors. In an immunocompetent host, it likely represents colonization [7]. In hospitalized neutropenic patients, it often indicates IPA with a positive predictive value as high as 90%; however, a negative sputum culture in an at-risk patient does not rule out IPA [8]. Bronchoalveolar lavage (BAL) culture may be helpful in the setting of diffuse lung involvement or focal disease if the associated bronchus is targeted for sampling; angioinvasive disease and diffuse lung involvement tends to be more common in severely neutropenic patients such as those with acute leukemia or leukocyte counts less than 100/mm³ [9]. While diagnostic yield is low with sensitivity around 50%, it has high specificity (97%) for IPA [8,10]. Given the low diagnostic yield and sensitivity of traditional respiratory cultures, laboratory techniques for Aspergillus spp. detection are being increasingly utilized.

2.1.2. Antigen Testing

Galactomannan and β-1,3-D glucan are cellular wall constituents of Aspergillus spp. that can be detected to aid in the diagnosis of IPA. Galactomannan is more specific to Aspergillus spp. than β-1,3-D glucan, which will be discussed in greater detail with Candida spp. and Pneumocystis. Galactomannan is released by Aspergillus spp. during replication and can be detected in body fluids using a double-sandwich Enzyme Immunoassay (EIA) (Table 1) [11]. Serum galactomannan is believed to have the best sensitivity in patients with severe neutropenia (<500 cells per mm³) [12]. Several hypotheses exist to potentially explain this observation: (1) fungal burden is higher in severely neutropenic patients, (2) galactomannan is consumed by the immune system in non-neutropenic patients, and (3) angioinvasive disease may be less severe in non-neutropenic patients [12]. The sensitivity of serum galactomannan in patients with hematologic malignancy and hematopoietic stem cell transplant is around 71% with a specificity of 89% [13]. The sensitivity is estimated to be around 25% in non-neutropenic patients, but may be increased to 88% by utilizing BAL samples to detect galactomannan in such populations [13].

Galactomannan is present in food items such as yogurt, cream cheese, and ice cream, and may be absorbed via the digestive tract in patients with mucositis [14]. Other causes of false-positive galactomannan assay are penicillin antibiotics or colonization/infection with other molds such as Penicillium, Fusarium, or Zygomycetes [15]. Secondary to false positive results and limited test reproducibility, it is recommended that positive serum galactomannan results be repeated [16]. Moreover, prophylactic antifungals targeting the cell wall such as triazoles may decrease the sensitivity of galactomannan; however, the degree of reduction has not been clearly established.

2.1.3. Aspergillus Polymerase Chain Reaction (PCR)

PCR-based techniques allow for detection and quantification of Aspergillus spp. DNA on serum and BAL samples (Table 1). Serum Aspergillus PCR has a reported sensitivity of 80.5% and specificity of 78.5% on single samples, but when 2 consecutive positive test results are required to define disease, the sensitivity decreases to 58.0% and specificity increases to 96.2% [17]. The sensitivity and specificity of Aspergillus PCR on BAL specimens is greater than 90% [18].

Prior studies have utilized different extraction techniques, amplification techniques, and primers; however, recent release of commercial assays has allowed for advances in the standardization of Aspergillus PCR [19]. Because of the advances in standardization of PCR techniques, the European Organization for Research and Treatment of Cancer (EORTC) now recognizes PCR as a diagnostic tool for IPA. Limitations with PCR techniques still exist [20]. Many of the prior studies have been conducted in patients with hematologic malignancies, and it may be difficult to distinguish airway colonization from infection [6].

2.2.1. Culture and Direct Visualization

Demonstration of a Mucorales spp. in tissue is required for definitive diagnosis of invasive Mucormycosis; however, issues of safety (e.g. bleeding risk, clinical stability, etc.) often preclude tissue sampling. Morphologically, species causing Mucormycosis appear as 6 to 25 µm hyphae featuring few or no septations and appearing ribbon-like with wide-branching angles [23]. The larger size and relative lack of septations allows for the rapid ability to distinguish Mucorales spp. from Aspergillus spp. [24]. Mucorales spp. grow within 1-7 days on most fungal culture media, but cultures may only be positive in 50% of cases, even when hyphae are seen on fungal smear secondary to the friable nature of the hyphae [23,25,26]. Mucorales spp. may not stain with Gram Stain and may be difficult to observe on KOH wet mount unless stained with chitin-binding stains such as Calcofluor [26,27]. Additionally, within tissues, Periodic Acid Schiff (PAS) or Gomori methenamine silver (GMS) stains are needed to easily visualize the organisms [26].

2.2.2. Molecular Strategies

No widely available or FDA approved molecular test for diagnosing Mucor exists. Molecular tests have been utilized to assist with Mucor species identification; however, this still requires deep tissue or invasive sampling [28]. More recently, serum PCR to detect free Mucor DNA has been successfully used to aid in diagnosis. [28]. While further study and standardization is necessary, this may provide a method for non-invasive testing.

3.1.1. Culture and Direct Visualization

The gold standard for diagnosing IC is identifying Candida spp. on blood or deep-tissue culture (Table 2), but the diagnostic yield of culture for all types of IC is low with a sensitivity of approximately 50% [6,31]. While blood cultures are almost universally positive in patients with active candidemia, they are positive in less than half of patients with candidemia complicated by deep-seated infection that persists after organism has been cleared from the blood and almost no patients with only deep-seated infection [31]. Time to culture positivity may also be prolonged delaying treatment initiation. Additionally, deep-tissue cultures require invasive procedures and may not be feasible in critically-ill patients at-risk for IC.

3.1.2. Antigen Testing

Most Candida antigens are rapidly cleared by the immune system limiting diagnostic detection. Mannan and β-1,3-D glucan are abundant cellular wall constituents that have been used as targets for diagnostic testing; however, only β-1,3-D glucan detection assay is approved by the US Food and Drug Administration (FDA) for clinical use in the United States (Table 2), while mannan and anti-mannan IgG tests are also available in Europe [31]. As with any antibody test, anti-mannan IgG assays are limited by the ability of the host to mount an appropriate immune response. Mannan and anti-mannan IgG assays have sensitivities less than 60%, but the specificity of each assay is close to 90% for IC [32].

β-1,3-D glucan is contained in the cell walls of most fungi with some notable exceptions including Cryptococcus spp. and Mucorales spp. The sensitivity of β-1,3-D glucan for IC is around 80% with a specificity of 60% [6]. β-1,3-D glucan assays, including Fungitell, do not differentiate Candida spp. from other fungi and can be falsely positive in a variety of scenarios including patients receiving hemodialysis, intravenous immunoglobulin, cardiopulmonary bypass, or blood transfusions [33]. β-1,3-D glucan may have role in ruling-out IC, but it should not be used as a stand-alone test to diagnose IC.

3.1.3. Candida Molecular Tests

T2Candida panel is an FDA approved test that detects Candida directly within the whole blood and can be used for candidemia (Table 2) [31]. The platform works to lyse red blood cells and concentrates Candida cells. The Candida DNA is then amplified and detected by T2 magnetic resonance [31]. Results are either positive or negative. T2Candida panel may remain positive for several days after candidemia has cleared, which may be useful in patients with hematogenous seeding and deep-seated infection [31].

3.2.1. Culture and Direct Visualization

Culture or direct visualization is the gold standard for diagnosis of infection with Cryptococcus spp. Cryptococcus spp. grows readily on Sabouraud dextrose agar at 30 degrees Celsius in aerobic conditions with growth typically noted within 7 days; however, treatment with antifungal therapy prior to obtaining culture specimens may delay growth (Table 2) [36]. Cryptococcus spp. appear as narrow-based budding yeast measuring 4-10µm. The yeast is surrounded by a characteristic capsule and may be visualized with multiple staining modalities [37].

Although pharmacologic management principles are similar, it is recommended that labs routinely distinguish between Cryptococcus spp. due to differing presentations and increased complications of Cryptococcus gattii, as well as Cryptococcus gattii association with outbreaks in temperate climates [38,39]. The traditional method to differentiate the two species via culture is use of Canavanine-glycine-bromothymol (CGB) agar as growth of Cryptococcus gattii produces a blue hue [37].

Among respiratory specimens, the source of specimen plays a role in the effectiveness of diagnosis. Bronchial washings or BAL are superior to sputum or biopsy for culture yield. In a retrospective review of 36 patients with Cryptococcal pulmonary infections in immunocompetent hosts, sputum culture was positive in 62.5% while cultures from bronchial washings and BAL were positive in 83.3% [40]. Smears from sputum and bronchial washing/BAL specimens were universally negative [40]. Among patients who had tissue specimens obtained via transbronchial biopsy, fine needle aspiration biopsy, or surgical lung biopsy, histopathology was positive in 76.9% and cultures positive in 72.7% [40].

3.2.2. Antigen Detection

Antigen detection provides a rapid method of evaluation for infection with Cryptococcus spp. Three methods of antigen detection are available: a latex agglutination assay, EIA, and lateral flow immunoassay, all of which detect components of the Cryptococcus capsule (Table 2). In addition to serum, Cryptococcal antigen may be tested for in other fluid specimens such as sputum, bronchial washings/BAL, pleural fluid, cerebrospinal fluid (CSF), and urine. While serum antigen testing demonstrates sensitivities and specificities greater than 90% with disseminated disease, testing performs more poorly in cases of isolated pulmonary infection where positive serum antigen tests have been reported in 73.9% of cases [41,42]. Negative serum antigen testing in the presence of pulmonary infection is thought to represent low extrapulmonary organism burden, or perhaps infection with a capsule-deficient strain. False positive tests have been associated with the presence of Rheumatoid factor or infections with Klebsiella pneumonia, Trichosporon beigelii, Stomatococcus mucilaginosus or Capnocytophaga canimorsus [35]. In cases where bronchoscopy is performed, Cryptococcal antigen testing on BAL samples provides excellent diagnostic utility with sensitivities and specificities as high as 100% and 98%, respectively [43]. Pleural effusion is an uncommon presentation of pulmonary Cryptococcus, and while data is limited, Cryptococcal antigen on pleural fluid has been shown to be positive in 80% of such cases [44].

5.1.1. Culture and Direct Visualization

Culture and direct visualization remain the gold standard for definitive diagnosis; though the time consuming and technically challenging nature of cultures are significant limitations (Table 3) [6]. Additionally, culture results in isolation are often unable to serve as the sole identifier of Histoplasmosis. A 2011 multicenter study of diagnostic tests for Histoplasmosis demonstrated culture sensitivity of 74% for disseminated disease, however, for isolated pulmonary Histoplasmosis, the sensitivity was 0%, 54%, and 67% for acute, subacute, and chronic disease, respectively [58]. Direct visualization is an alternative to culture. Histoplasma appears as ovoid yeast cells measuring 2 µm to 4 µm with budding yeast connected at a narrow base [59].

5.1.2. Antigen Testing

Testing for Histoplasma antigens has evolved as a method to aid in both the accuracy and rapidity of diagnosis. Updated techniques allow for quantitative methodology increasing diagnostic sensitivity (Table 3). A 2016 meta-analysis evaluating antigen testing on serum and urine specimens demonstrated an overall sensitivity of 81.4% and specificity of 98.3% among combined cases of disseminated Histoplasmosis and pulmonary Histoplasmosis [60]. Sensitivity of antigenemia was slightly more favorable than antigenuria, 83.9% vs 79.5% [60]. Many of the patients included in this meta-analysis had disseminated disease. In cases of isolated pulmonary Histoplasmosis, sensitivity of antigenemia and antigenuria are lower at 63.3% and 42.7% respectively and 67.5% when used in combination [61]. Antigen detection on BAL has excellent sensitivity (93.5%) and specificity (97.8%) [62].

A major limitation of Histoplasma antigen testing is cross-reactivity with other endemic mycoses. A high degree of cross-reactivity is seen with Blastomycosis, and less frequently with Coccidioides, Paracoccidioides, African Histoplasma, and Talaromyces and rarely with Aspergillus [58,59,62]. No cross-reactivity has been found with Candida or Cryptococcus [63].

5.1.3. Antibody Testing

Antibody testing for Histoplasma utilizes multiple methods—Complement Fixation (CF), Immunodiffusion (ID), and EIA. CF and ID are the most commonly used. ID identifies precipitating antibodies designated M or H bands, and has a lower sensitivity than CF but a higher specificity (Table 3) [64].

More recently, an EIA has been developed as a method to quantitatively measure serum levels of IgM and IgG antibodies. This assay demonstrates a sensitivity of 67.5% and 87.5% and specificity of 97.0% and 95.0% for IgM and IgG, respectively; however, when tested in combination, sensitivity and specificity were shown to be 88.8% and 91.9%. When directly measured against both CF and ID, sensitivity was greater for EIA (73.1% for CF and 55% for ID) [61]. A potential limitation of EIA is cross-reactivity with Blastomyces and Coccidioides, though cross-reactivity has been shown to be less than antigen testing [58].

Like other serologic testing methods, the host immune response influences accuracy of results. While the sensitivity of antibody testing in immunocompetent hosts may approach 95%, antibody testing in HIV-AIDS patients and solid organ transplant recipients is less than 50% and 20%, respectively [58,65]. Additionally, the time course of the infectious syndrome may influence the yield as serologic testing within the first month of infection may have sensitivity as low as 67% [24,58].

5.1.4. Histoplasma PCR

Real-time PCR for Histoplasma has been utilized on tissue, serum, urine, and respiratory specimens [66]. When utilizing culture as the reference standard, PCR appears to have a sensitivity of greater than 95% [66]. For patients with severe immune compromise, PCR may be an adequate test as organism burden is likely to be high, and cultures are likely to be positive. When compared to antigen testing, PCR is positive in less than 10% of disseminated Histoplasma cases when antigen testing is positive [67]. Antigen testing appears to be more sensitive than PCR for disseminated Histoplasma.

5.2.1. Culture and Direct Visualization

Fungal culture is the gold standard for diagnosis of Blastomycosis, and cultures can be performed on a variety of specimens (Table 3). Sputum, tracheal secretions, and gastric washings are less-invasive methods of obtaining culture specimens. Culture sensitivity is 75% on a single sample but may be increased to 86% if multiple samples are collected. More invasive methods, such as bronchoscopy or surgical sampling, may yield sensitivities as high as 100% [68]. A significant limitation of culture is the long duration of time required for positive results, which may take as long as 5 weeks [6].

Visualization of Blastomyces allows for rapid identification; however, the sensitivity is only 36-46% depending on the number of specimens evaluated [68]. The classic appearance of Blastomyces is a round or ovular multinucleated yeast cell measuring 8 to 15 µm with a single broad-based bud [69]. Visibility may be enhanced with KOH preparation to degrade tissue with or without Calcofluor white; The organism may also be visualized on PAS and GMS stains [69].

5.2.2. Antigen Testing

Antigen testing provides a method for rapid identification in cases of suspected Blastomycosis (Table 3). Urine Blastomyces antigen testing has an overall sensitivity of 76%−93%; however, specificity is only around 80% due to high cross-reactivity with Histoplasma (96.3%), Paracoccidioides (100%), and Talaromyces (70%) [70–72]. Cross-reactivity is present to a far lesser degree in patients with Cryptococcus (2.9%) and Aspergillus (1.1%), and no cross-reactivity is noted with Coccidioides or Candida spp. [70]. Serum and BAL antigen testing have sensitivities around 55% and 63%, respectively [72].

5.2.3. Antibody Testing

Antibody testing provides an additional route for rapid diagnosis of Blastomycosis; although the literature regarding its utility is more limited compared to antigen testing (Table 3). Testing is currently available via EIA, ID, or CF for antibodies to the A antigen of Blastomyces and sensitivities are low to moderate: 8% for CF, 27% for ID, and 77% for EIA [73].

Recently a new EIA for antibodies to the Blastomyces surface protein, BAD-1 has been developed, which shows more promise than traditional testing targeted at the A antigen. Sensitivity of this test is 87.8%, and false positives in patients with Histoplasma are around 6% [74].

5.3.1. Culture and Direct Visualization

Culture or direct visualization represents the gold standard for diagnosis of Coccidioidomycosis. Visually, Coccidioides spp. appear as spherules ranging from 20 to 200 µm and filled with 2 to 4 µm endospores (Table 3) [75]. Though multiple stains are of utility and the appearance of Coccidioides spp. is quite specific, direct visualization is positive in less than one-third of, and cultures are positive in around half of immunocompromised patients [6].

5.3.2. Antigen Testing

The literature is scarce, but antigen testing in urine and serum has been studied and may provide rapid diagnosis of Coccidioidomycosis. In a small study testing urine for the presence of Coccidioides antigen, sensitivity was 70.8% (Table 3) [76]. The majority of these patients were immunocompromised, and many had severe disease which may limit the generalizability [76]. There was 58% cross-reactivity with Histoplasma antigen testing, and testing was negative in 99.4% of healthy controls [76]. Serum antigen testing is another possibility to aid in diagnosis; however, the studied sample sizes are low with reported sensitivities ranging between 55.6-77.8% [77–79].

5.3.3. Antibody Testing

Antibody testing is frequently used with EIA, ID, and CF available as testing methods. While EIA and ID are used as qualitative methods in the diagnosis of Coccidioidomycosis, CF is utilized as a quantitative method. Higher titer levels with CF testing correlate with disseminated disease and allow for monitoring treatment response (Table 3) [80]. While a wide variety of literature is available regarding such tests, a limitation of the literature is the changing methodologies of testing, and infrequent use of culture or direct visualization as the gold standard [6]. Similar to other serologic methods, timing of testing in relation to infection onset and the underlying immune function may impact results.

A recent 2017 study of IgM and IgG antibody testing via EIA in 103 patients with Coccidioidomycosis showed sensitivity of 61.2% for IgM and 87.4% for IgG with a combined sensitivity of 88.3% [81]. When evaluating cases of isolated pulmonary infection, sensitivity was slightly higher with 66.7% for IgM and 90.2% for IgG and a combined sensitivity of 90.2% [81]. Just under one-third of studied patients were immunocompromised, and while sensitivities of EIA were reduced in this population, it was not statistically significant [81].